BACKGROUND: Though cooperative trial groups use different treatment protocols for newly diagnosed childhood acute lymphoblastic leukemia (ALL), all achieve high cure rates. The healthcare utilization and costs associated with different treatment strategies have not been rigorously compared. Minimizing utilization and costs may increase quality of life and decrease health system burden. We compared Children's Oncology Group (COG) and Dana-Farber Cancer Institute (DFCI)-based treatment.

METHODS: We identified all children diagnosed with ALL in pediatric cancer centers in Ontario, Canada between 2002 and 2012 through the Pediatric Oncology Group of Ontario Networked Information System (POGONIS), a provincial pediatric cancer registry. Detailed data on demographics, disease risk factors (e.g. cytogenetics, minimal residual disease), treatment (e.g. treatment protocol, start and end date of each therapy phase) and events (relapse/progression, death, second cancer) were captured via chart abstraction. Treatment protocols were categorized as either based on COG or DFCI trials. Linkage to population-based health services databases identified all outpatient and emergency department (ED) visits, hospitalizations, and physician billings. Healthcare utilization-associated costs were determined through validated costing algorithms. Chemotherapy-associated costs were calculated separately using local prices. All administered doses of asparaginase (ASNase), including E. Coli, PEG-ASNase, and Erwinia ASNase were recorded. Event-free survival (EFS), overall survival (OS), healthcare utilization rates, and costs were compared between COG and DFCI-treated patients while adjusting for demographics and disease-factors using appropriate regression models. Healthcare-associated costs, ASNase costs, and total chemotherapy costs (2018 Canadian dollars) were compared.

RESULTS: The study cohort included 802 patients, 146 (18.2%) of whom were treated on DFCI-based protocols. Median follow-up did not differ between between COG and DFCI patients; nor did EFS or OS. When adjusted for all demographic and disease-related variables, COG patients experienced significantly higher rates of ED visits [rate ratio (RR) 1.3, 95% confidence interval (95CI) 1.1-1.5; p=0.01]. Neither hospitalization rates nor rates of inpatient days differed between the two groups of patients. However, rates of outpatient visits were 60% higher among DFCI patients (RR 1.6, 95CI 1.5-1.7; p<0.0001).

The median healthcare-associated cost in the first 5 years following initial diagnosis was $193,700 among COG patients [interquartile range (IQR) 149,200-272,700] compared to $288,000 among DFCI patients (IQR 233,300-407,300; p<0.01), mainly attributable to the cost associated with outpatient visits. In adjusted analyses, DFCI-associated costs were 70% higher (RR 1.7, 95CI 1.5-1.9; p<0.0001).

The median ASNase-related cost was similar between COG and DFCI patients [$21,100, IQR 14,800-35,400 vs. $19,900, IQR 15,600-39,900; p=0.91]. The median total chemotherapy cost was higher among COG patients ($29,100, IQR 20,300-50,300 vs. $22,400, IQR 17,300-42,600; p<0.001]. However, ASNase and total chemotherapy costs were highest in DFCI patients treated with PEG-ASNase instead of E. Coli ASNase (N=36), reflecting contemporary practice. Among such patients, median ASNase-related costs were $45,200 (IQR 7,000-68,500) and median total chemotherapy costs were $48,000 (IQR 11,000-73,400) (p=0.004 and p=0.09 vs. COG patients).

CONCLUSIONS: Though COG and DFCI protocols are associated with equivalent EFS and OS, patterns of healthcare utilization differ with the former associated with a 30% increase in the rate of ED visits and the latter associated with a 60% increase in the rate of outpatient visits. Overall, healthcare utilization-associated costs were increased in DFCI-treated patients. Though ASNase costs historically did not differ, the shift to PEG-ASNase is associated with higher ASNase and total chemotherapy costs on DFCI protocols. Decreases in PEG-ASNase cost and the ability to administer intravenous or intramuscular chemotherapy at home would decrease overall healthcare utilization and costs, and mitigate differences between COG and DFCI protocols. These results can inform efforts to decrease burden on both families and health systems.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.